The goal of this research is to find out how integrated systems of knowledge are acquired over several years and are maintained, or decline over several decades. This will contribute new information about long-term memory by enlarging the data base of memory research from the current time-compressed longitudinal laboratory investigations of retention over short time spans, to knowledge systems acquired in real-life learning situations and tested by a cross-sectional approach 1-50 years after the knowledge system was acquired. The cross sectional approach sacrifices laboratory control over conditions of original learning and rehearsal of the material and compensates for this loss by an adjustment technique based upon multiple regression calculations. The adjusted memory data then reflect performance estimates under constant learning and rehearsal conditions for all time groups to be compared. A further objective of this research is the investigation of rehearsal conditions necessary to maintain knowledge systems at various levels of integrity. These findings can yield guidelines relevant to educators, scientists and others interested in finding out under what circumstances systems of knowledge once acquired can be maintained intact.